Mistral AI
Mistral AI
Mistral AI develops frontier large language models and generative AI systems for organizations building intelligent applications and agents. Its product portfolio spans Vibe for AI-assisted long-horizon work and coding, Studio for creating and operating agents and applications, Forge for training and evaluating custom models, and Compute for model training and inference infrastructure. The company also provides APIs, specialized models, enterprise customization, deployment services, and industry solutions across areas such as finance, the public sector, manufacturing, energy and utilities, and transportation. Mistral AI supports both cloud-hosted and self-hosted deployments, with a focus on privacy-conscious enterprise use and tailored AI solutions.

LLM Scientist, AI for Science — Mistral AI, Paris Hybrid

Develop large language models and agentic AI for Mistral AI’s scientific discovery solutions. Partner with experts across biology, chemistry, physics, and materials science to advance practical AI applications.

Description

  • Build and optimize large language models, agentic AI systems, and deep learning frameworks for scientific discovery
  • Advance LLM performance in literature synthesis, autonomous experimentation, and rigorous derivations
  • Support end-to-end training of scalable, generalizable LLM-based solutions for scientific use cases
  • Create scalable workflows for collecting high-quality scientific data through synthetic and crowdsourced pipelines
  • Apply and advance reinforcement learning and agentic systems for challenges in the natural sciences
  • Work with domain experts to identify AI opportunities and refine models using their feedback
  • Help develop infrastructure for scientific model training and open-source AI tools for science
  • Define research priorities for AI in science while investigating high-risk, high-reward ideas
  • Turn advances in AI research into measurable scientific impact

Requirements

  • Strong expertise in large language models, deep learning, or agentic AI
  • Demonstrated success applying AI or machine learning to real-world problems through publications, open-source work, or deployed systems
  • Extensive computer science experience
  • Ideally, experience scaling synthetic data generation
  • Working knowledge of current LLM architectures and post-training methods
  • Hands-on experience with PyTorch, TensorFlow, or JAX
  • Experience with model optimization and scaling methods, including fine-tuning and RLHF
  • Ability to understand at least one natural science domain at a high level
  • Graduate-level specialization and broad undergraduate knowledge in a natural science field
  • Proficiency in scientific computing, numerical methods, and high-performance computing
  • Strong software engineering practices
  • Ability to explain technical AI concepts to non-specialists
  • Collaborative, low-ego approach with a willingness to handle operational challenges
  • Availability to work from the Paris office approximately three days per week

Benefits

  • Healthcare coverage
  • Parental leave
  • Retirement plans
  • Relocation support
  • Wellness programs
  • Meal allowances
  • Transportation allowances
  • Other location-specific perks

Related Jobs

CONVACT - Agentur für Suchmaschinenoptimierung

Remote Sales Representative – SEO and Google Ads

CONVACT - Agentur für Suchmaschinenoptimierung
1 – 10 Employees
B2BMarketing

Follow up with leads and cold-call prospective clients for Convact’s SEO and Google Ads agency. Book appointments with online shops and e-commerce businesses across the DACH region.

Open